Four-Channel DC Arc Fault Detector for 100 kW Combiner Box

Since the announcement of UL 1699B in 2018, several studies on DC arc detection methods have been conducted to meet this standard. This new standard necessitates tests of DC arc detection and unwanted tripping that are caused by maximum power point tracking, DC switching surge, etc. The frequency ba...

Full description

Saved in:
Bibliographic Details
Main Authors: Lee, Jin Han, Lee, Jin, Kim, Yong Joo, Ahn, Jae Bum, Ryoo, Hong Je
Format: Conference Proceeding
Language:English
Subjects:
AFD
DC arc
Discrete wavelet transforms
DWT
Fault detection
Four-channel
PV system
Software
Software algorithms
Switches
Switching frequency
Transforms
UL 1699B
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Since the announcement of UL 1699B in 2018, several studies on DC arc detection methods have been conducted to meet this standard. This new standard necessitates tests of DC arc detection and unwanted tripping that are caused by maximum power point tracking, DC switching surge, etc. The frequency band of the DC arc signal is generally below 100 kHz. Unfortunately, this frequency band overlaps with an inverter switching frequency. In addition, burst pulse-like arc signals appear for less than 20 msec and then fade away. The laboratory experiment on a four-channel DC arc fault detector (AFD) conducted in this study introduces its specific design configuration and DC arc detection. A STM32F4 microcontroller unit (MCU) and discrete wavelet transform (mother wavelet: Daubechies 4) algorithm are adopted as hardware and software, respectively, to detect DC arc faults. Once 4 AFD units with 16 DC arc sensors (4 sensors/AFD unit) are installed into the 100kW combiner box, the 4 AFD units must complete one arc monitoring cycle within 20 msec. Based on the laboratory study, both time reduction in the arc detection process and upgrade of the MCU to STM32F7 are proposed to overcome the limited application of the four-channel DC AFD.
ISSN:2644-271X
DOI:10.23919/CMD54214.2022.9991290